室内环境监测中的BP神经网络算法的改进及其仿真分析

发布时间：2018-06-30 10:36

本文选题：物联网 + 居家室内环境　；参考：《南京邮电大学》2017年硕士论文

【摘要】：随着信息感知技术、人工智能技术和网络技术的发展,人们生活质量得到了很大的提高。由于室内环境作为人们主要的生活环境,居家室内环境监测系统在这背景下应运而生。如今的居家室内环境监测系统相对于传统的居家环境监测系统提出更多的要求,要求具有对室内环境数据采集能力,对海量数据的处理能力,以及决策监测能力,能够为用户提供智慧化、个性化的监测服务。本文在神经网络、嵌入式技术、ZigBee无线技术等相关技术的基础上,提出了一种改进型BP神经网络算法并应用于居家室内环境监测系统。本文首先介绍了现有的居家室内环境监测系统国内外研究现状,分析了当前系统的局限性,介绍了系统中使用的关键技术,分别是神经网络技术、无线传输技术和嵌入式技术。接着设计并实现了一种基于改进型BP神经网络的居家室内环境监测系统,系统主要分为应用层、网络层、感知层和控制层。本文针对标准BP神经网络误差收敛速度缓慢,改进了BP神经网络的学习率自适应算法,然后把改进型BP神经网络算法应用到居家室内环境监测系统中,算法采用MATLAB进行仿真,对比使用标准BP神经网络和改进型BP神经网络的均方误差收敛曲线,得出了改进型BP神经网络算法的误差收敛速度显著提高。最后,对该系统进行软硬件调试,展示改进型BP神经网络算法在居家室内环境监测系统中应用的可行性。
[Abstract]:With the development of information perception technology, artificial intelligence technology and network technology, people's quality of life has been greatly improved. Because the indoor environment is the main living environment, the indoor environment monitoring system comes into being under this background. Compared with the traditional indoor environment monitoring system, today's indoor environment monitoring system has more requirements, such as the ability of collecting indoor environment data, the ability of processing massive data, and the ability of decision-making monitoring. To provide users with intelligent, personalized monitoring services. On the basis of neural network, embedded technology and ZigBee wireless technology, this paper presents an improved BP neural network algorithm and applies it to the indoor environment monitoring system. This paper first introduces the current research status of the existing indoor environment monitoring system at home and abroad, analyzes the limitations of the current system, and introduces the key technologies used in the system, which are neural network technology, wireless transmission technology and embedded technology. Then a home indoor environment monitoring system based on improved BP neural network is designed and implemented. The system is divided into application layer, network layer, perceptual layer and control layer. Aiming at the slow convergence speed of the standard BP neural network error, this paper improves the learning rate adaptive algorithm of the BP neural network, and then applies the improved BP neural network algorithm to the indoor environment monitoring system. The algorithm is simulated by MATLAB. Compared with the mean square error convergence curve of standard BP neural network and improved BP neural network, the speed of error convergence of the improved BP neural network algorithm is improved significantly. Finally, the software and hardware of the system are debugged, and the feasibility of applying the improved BP neural network algorithm to the indoor environment monitoring system is demonstrated.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP183;TP274

【相似文献】